Bangor Business School
Working Paper
BBSWP/13/005
COMPETITION AND BANK RISK
THE ROLE OF SECURITIZATION AND BANK CAPITAL
By
Yener Altunbas
Michiel van Leuvensteijn
David Marques-Ibanez
October 2013
Bangor Business School
Bangor University
Hen Goleg
College Road
Bangor
Gwynedd LL57 2DG
United Kingdom
TEL: +44 (0) 1248 382277
COMPETITION AND BANK RISK
THE ROLE OF SECURITIZATION AND BANK CAPITAL
Yener Altunbas1
Centre for Banking and Financial Studies, Bangor University
(Corresponding author)
Michiel van Leuvensteijn
APG, Corporate Strategy and Policy Department
David Marques-Ibanez
European Central Bank, Research Directorate
Abstract: We find that the increased use of securitization activity in the banking sector augmented the
effect of competition on realized bank risk during the 2007-2009 crisis. Our results suggest that
securitization by itself does not lead to augmented risk while higher levels of capital do not buffer the
impact of competition on realized risk. It follows that cooperation between supervisory and competition
authorities would be beneficial when acting on the financial stability implications of financial innovation
and the effects of bank capital regulation.
Keywords: securitization; competition; bank risk
JEL classification: G21; D22
1 Corresponding author: Yener Altunbas, Centre for Banking and Financial Studies, Bangor University, Bangor, Gwynedd, LL57 2DG, UK, Phone:
+441248382191. Email: [email protected]. Disclaimer: The opinions expressed in this paper are those of the authors only and do not necessarily reflect those of the European Central Bank (ECB) or APG. We are grateful to Leonardo Gambacorta, Simone Manganelli, Alex Popov and participants at seminars at
the European Central Bank and Bangor University for useful comments.
I. Introduction
The 2007-2009 financial crisis brought the connection between competition and bank stability again to
the fore. Historically this was not always the case. Indeed, in the five decades following the Great
Depression, the banking industry was tightly regulated and, as a result, competition was largely contained
in most developed countries (Benston, 1998). Essentially the underlying idea at the time was that
competition was detrimental to financial stability. In contrast, in the three decades prior to the 2007-2009
financial crisis, an unprecedented process of deregulation took place in the banking sector (Vives, 2001).
Broadly speaking, during this latter period, efficiency considerations took a preceding role over financial
stability concerns as the deregulatory process aimed primarily at improving the efficiency of the financial
system.
The economic theory is however unclear on the expected effect of increased competition on bank
risk.2 One strand of the literature argues that competition endangers financial stability. In a nutshell, it
suggests that increased competition for deposits would erode profits lowering the market power of banks
thereby depressing their charter value. This decline coupled with the existence of limited liability (and a
‘quasi’ flat-rate deposit insurance in most countries) would encourage banks’ owners to expand and take
on new risks by shifting the risks to depositors and ultimately the government (Keeley, 1990; Matutes and
Vives, 1996; Hellmann, et al., 2000). In this direction also, more concentrated (and possibly also less
competitive) markets have larger banks that might be better able to diversify and therefore contribute to
more stable financial system than less concentrated markets (Ramakrishnan and Thakor, 1984; Boyd and
Prescott, 1986;Williamson, 1986; Allen, 1990).3
In contrast, there is a parallel strand of the literature that argues that competition among banks
would enhance, rather than disrupt, financial stability. This latter strand of papers considers the effect of
bank competition on the loan market as opposed to focusing on the effect of bank competition on the
deposit market only. The additional consideration is that competition would lower the lending rate
charged to borrowers thereby raising their profits so they would have fewer incentives to take on new
risks lowering, as a result, the overall risk profile of the bank (Boyd and De Nicoló, 2005).4
2 For very good overviews see for instance Vives (2010), Carletti and Hartmann (2003), Cetorelli (2001). 3 A related argument also supporting this positive effect of competition on risk refers to the number of banks to be supervised by the authorities. A more concentrated banking system implies a smaller number of banks. This would be expected to reduce the supervisory burden and improve the overall quality of
the supervision enhancing, as a result, the stability of the banking system (Allen and Gale, 2000). 4 Martinez-Miera and Repullo (2005) show that this argument does not account the fact that lower rates also reduce the banks’ revenues from non-defaulting loans.
Comparing different models, Allen and Gale (2004) illustrate how crucial the specific details of the
models are in concluding whether competition leads to more or less financial stability suggesting that the
issue is multifaceted. There is also a vast empirical literature analyzing the relationship between
competition and bank risk which tends to provide similarly ambiguous results.5 Hence further evidence on
the different factors at work underlying this relationship, both at the empirical and theoretical level seems
warranted. Building on the existing literature, we focus on two main variables (securitization and bank
capital) that might have affected this relationship in recent years.
Turning to the first factor, large increases in the use of securitization activity in the decades before
the crisis represented a major structural development in the banking sector in most developed countries
(Marques-Ibanez and Scheicher, 2010).6 Securitization allowed banks to turn traditionally illiquid claims
(overwhelmingly in the form of bank loans) into marketable securities and sell them off to the financial
markets. In principle, from the perspective of individual banks, securitization allowed banks to diversify
their risk portfolio more effectively, both geographically and by sector. Yet banks might also respond to
the static reduction in risks due to securitization by taking on new ones.7
Our first hypothesis acknowledges that securitization activity might have impacted on the effect of
competition on bank risk in recent years. Scant empirical evidence from the pre-crisis period suggests that
banks that were more active in the securitization market were found to have higher profitability, and were
often better capitalized (Cebenoyan and Strahan, 2004). At the same time, in the long-run banks in more
competitive markets with high levels of securitization activity might be expected to have riskier loan
portfolios (Ahn and Breton, 2011).8 More broadly, as competition increases banks might also respond to
the static reduction in risks due to securitization by taking on new ones, in particular, by loosening their
lending standards, increasing their leverage, or becoming systemically riskier (Mian and Sufi, 2009; Keys
et al., 2010; Nijskens and Wagner, 2011).9 There is, however, some recent and growing empirical
evidence suggesting that there are no differences in performance (i.e. credit risk) between equivalent
securitized and non-securitized loans (Albertazzi et al., 2011, and Benmelech et al., 2012).
Our second hypothesis recognizes that in the run-up to the 2007-2009 crisis, prudential regulators
responded to the process of financial de-regulation by giving bank capital a central position as a
5 See Beck (2008) and Vives (2010) for two excellent surveys of this large literature. 6 While securitisation in a narrow view has been used as a technique in the United States for more than fifty years, the decade prior to the financial crisis
coincided with spectacular increases in the amount of securitisation activity, in the number countries using these techniques and in the development of new credit risk transfer instruments. 7 See Gorton and Metrick (2012) for a lucid and comprehensive recent review of this literature. 8 In this model securitization works as a signaling device indicating that banks will reduce monitoring. 9 A number of recent theoretical models illustrative the social costs of securitization. See for instance Stein (2011) or Gorton and Ordoñez (2012).
supervisory tool influencing banks’ behavior. As a result, in most recent models of competition bank
capital plays a major role. Higher levels of bank capital would raise banks’ charter value thereby
buffering risk-taking incentives in competitive markets. Hence banks with high capital ratios in more
competitive markets would reduce bank risk by more than banks with high capital ratios in less
competitive markets (Keeley, 1990; Hellman et al., 2000). Equity also provides banks with greater
incentives for exerting better monitoring of borrowers as it forces banks to internalize the costs of their
default.10
At the same time, there are also reasons to expect a positive relationship between capital and
risk particularly as competition increases. In fact the theoretical literature gives us grounds for doubting
that increased capital requirements will necessarily result in reduced risk-taking incentives (Gale and
Özgür, 2005; VanHoose, 2007). For instance, an increase in the required capital ratio can force banks to
take on more risk in order to achieve target rates of return (see Gale, 2010). Empirically also, higher
levels of capital may simply be the result of regulators forcing riskier banks to hold higher buffers. There
is, in fact, significant evidence finding a positive relationship between higher levels of bank capital and
risk (see for instance Berger and Bouwman, 2013; Delis and Staikouras, 2011).11
II. MODEL, IDENTIFICATION STRATEGY AND DATA RESULTS
A major challenge of the empirical literature analyzing the relationship between competition and risk is
when to time the realization of bank risk (Beck, 2008) as there is an important lag between the period in
which risk-taking takes place and its materialization. Another, important modeling challenge is that an
important component of bank risk often materializes only in the event of a crisis.
We exploit the materialization of bank risk during the 2007-2009 crisis and consider whether the
ex-ante cross-sectional variability in bank characteristics and competitive conditions prior to the crisis are
related to the ex-post materialization of bank risk.12
Our approach assumes that to a large extent the
measurement of risk can only be gauged when an extreme event materializes. That is, when a crisis
occurs.
Our model measures the probability of a bank belonging to the group of riskier institutions during
the 2007-2009 crisis. We define as “risky” as those banks that received direct public assistance during the
crisis and creating a binary variable (RISKY) that takes the value of 1 if the bank had any form of
10
There is evidence from the crisis suggesting that higher capital might be associated to more risk (Demirguc-Kunt et al. (2010). 11 In fact many of the banks failing during the crisis had capital levels above the average of their peers (Haldane and Madouros, 2012). 12 Other studies have used a similar strategy to consider how certain pre-crisis characteristics were linked to stock market performance during the recent crisis (see for instance Beltratti and Stulz, 2012).
financial support during the crisis after 2007Q2-2008Q4 period and 0 otherwise. Hence our definition of
risky banks refers only to the risks that materialized during the crisis. The statistical sources used and a
brief description of the main variables included in our study are provided in the first two columns of
Table I. In total we have 495 bank observations for the largest listed bank parent companies in Austria,
Belgium, the Netherlands, Germany, France, the UK, Italy, Portugal, Spain and the United States.13
For each country, we construct a measure of competition and based on the ex-ante literature, we
also add a number of other factors likely to impact on bank risk (see Altunbas et al., 2011). Hence the
vector X also includes 5 bank-specific characteristics: securitization activity (SEC), capital-to-assets
(CAP), size in terms of total assets (SIZE), loan growth (EXLEND) and, short-term deposits (DEP). More
importantly for our purposes, we include the interaction between competition (COMP) and two main
bank specific characteristics: securitization (SEC) and bank capital (CAP). In this way, we assess whether
banks with different characteristics in terms of financial innovation and capital position adopted different
risk strategies in connection with the existence of high competition in the loan market (COMP*X).
The baseline empirical model is given by the following probit equation:
1 ( ' ' * )ikP risky X COMP Y X COMP X (1)
where P is the probability, Φ is the standard cumulative normal probability distribution, Y is a
vector of country dummies k where bank i has its main seat, and X a vector of bank-specific
characteristics of the same bank i over the five years prior to the crisis (2002Q2–2007Q2). This approach
limits endogeneity problems. The probit model is estimated by maximum likelihood.
Our measure of bank risk, RISKY, captures whether an institution received government support.
The construction of this variable is based on the collection of information related to the public rescue of
individual banks via capital injections or other government-sponsored programmes. Information on this
variable has been obtained from public and confidential sources including the European Commission,
European Central Bank (ECB), Bank for International Settlements (BIS), Bloomberg and the websites of
a number of governmental institutions and national central banks. Around 30% of the banks in our sample
needed financial support after the start of the crisis in August 2007.
13 In order to have a comparable sample, we focus on the parent company of the largest privately owned, listed commercial banks headquartered in those
countries. This reduces our sample from 1,100 initial institutions (often belonging to a larger group as branches or subsidiaries) to 495 banks. For a full description of the characteristics of the database and variable definitions see Altunbas et al. (2011).
In order to measure bank competition, we use the so-called Boone-indicator (Boone, 2008). This
measure of competition is based on the notion that in a competitive market, more efficient companies are
likely to gain larger market shares than in a non-competitive market. Compared to other measures of
competition like the Hirschman-Herfindahl index (HHI), the adjusted Lerner index, or the price cost
margin (PCM), the Boone indicator is monotone in intensity to competition and does not depend on the
Cournot model assumptions. For instance, the standard intuition of the HHI is based on a Cournot model
with symmetric firms, where a fall in entry barriers reduces the HHI. However, in a setting in which firms
differ in efficiency, an increase in competition reallocates output to the more efficient firms that already
had higher output levels. Hence, the increase in competition raises the HHI, but decreases the Boone-
indicator because of the strengthening relationship between performance and efficiency. Other commonly
used measures of competition such as the price-cost margin (PCM), or the so called Lerner index, have
similar disadvantages.14
Finally, heavier competition reduces the PCM of all firms. But since more efficient firms may
have a higher PCM (skimming off part of the profits stemming from their efficiency lead), the increase in
their market share may raise the industry’s average PCM, contrary to common expectations. As such, the
estimates of the PCM will typically underestimate the price-cost margin (PCM) and the actual level of
competition. For all these reasons, we chose the Boone-indicator as our preferred measure of competition.
We used the Boone indicator with country-level data as described in Van Leuvensteijn et al.
(2011, 2013). The Boone indicator varies from 0.2 to 5.6 and shows that, according to our metric,
Portugal appears to be one of the less competitive countries. The securitization variable, SEC reflects
averages of securitization flows ranging from 0 to 9.8% of total assets (source DCM Analytics Dealogic)
while the average capital to total assets (CAP) before the crisis around 9%. We compute a bank-specific
measure for credit expansion, EXLEND, by subtracting from each bank’s lending growth the average
expansion in bank lending for the whole banking industry in that country. The average short-term deposit
as percentage of assets, DEP, amounts to around 70%. As a macro variable we have included year on year
real GDP growth.
14 Graddy (1995) and Genesove and Mullin (1998) estimate the elasticity-adjusted Lerner index, Lerner-index times the price-elasticity of sugar demand and
show that the conjectural variation parameter can be interpreted as a measure of competition. Corts (1999) criticizes this approach suggesting that, in general, efficient collusion cannot be distinguished from Cournot competition using the elasticity adjusted Lerner index.
III. RESULTS
Table II, column I shows that the use of securitization (SEC), diminishes the likelihood of receiving
public assistance. In other words, off-loading risks to third parties via securitization by itself, did not seem
to have made banks more likely to default during the crisis period. In this direction also, better capitalized
banks (CAP) and those institutions more reliant on stable funding via customers deposits have lower
likelihood of default. The latter is not surprising as many whole sale markets dried up during the crisis
compared with traditional deposit funding which remained far more stable for most institutions. These
outcomes (columns II-VII) do not change when different specifications are introduced although the direct
impact of competition loses, of course, it significance when country dummies are introduced or when the
interaction between COMP and CAP are included.
Turning to our main hypotheses on the interaction terms of competition, our results show that
securitization compounds the direct effect of competition on bank risk. In other words, as competition
increases, banks with higher levels of securitization activity (SEC) also tended to have a riskier profile as
the crisis erupted (columns IV and V). One possible plausible explanation for these results is that
enhanced competition before the crisis might have lead to a higher resource to securitization activities
which eventually resulted in an increased risk profile by banks presumably due to reduced monitoring
incentives by those banks. In other words, it seems that securitization by itself (i.e. ‘per se’) is not
deleterious from a financial stability perspective but rather the interaction between competition and
securitization that potentially can pose a financial stability problem.
We also find that as competition increases, bank capital is not effective in counterbalancing the
direct effect of competition on financial stability. In other words, as competition intensifies, higher levels
of capital are not enough to support financial stability. At the same time, we also find that capital by itself
was effective in buttressing individual banks’ soundness during the crisis.
An important consideration is that as the variable competition was interacted (with securitization
and bank capital) in a non-linear model, the direct interpretation of the interaction variable could be
misleading (Ai and Norton, 2003). For instance, we found a positive estimated marginal effect for the
interaction term between competition and capitalization which implies that higher securitization would
make the positive effect of competition larger. This, however, does not mean that the interaction between
these two variables is always positive and siginificant across the whole sample.
Chart I presents two charts showing both the correct interaction effects as well as the incorrect
marginal interaction effects of our model for the variables interacting competition and securitization
(Panel A) and competition and capital (Panel B). The charts indicate that the interpretation of the
interaction terms is more complex than as suggested by the single marginal effect estimator calculated for
the whole sample as presented in Table II. The thrust of the interpretation, however, remains as both
panels suggest that only for a small percentage of the population the interaction effects are negative. That
is, Charts I.A and I.B show that for the overwhelming majority of the banks in our sample the values of
the interaction term are positive. This suggests that in most cases higher competition coupled with more
resource to securitization activity (or higher capital levels) is conducive to more bank risk. The
significance of the interaction terms is assessed in Chart II. It estimates the interaction effects for whole
sample and shows that these interactions are highly significant (i.e. 1%) and positive for a large group of
observations. More concretely, it is positively significant for nearly for 83% of the sample for the variable
interacting competition and securitization and 64% the variables interacting competition with bank
capitalization. Finally, as suggested by Green (2010), we have further represented graphically the
interaction variables according to the strength of competition and our results shows that the strength of
the interaction results applies in particular for those markets with higher levels of competition (i.e.
competition above the median).15
Finally our results are consistent to the use of other (continuous)
variables accounting for bank risk.16
IV. CONCLUSION
We found that as competition increases banks resorting more heavily to securitization activity have more
incentives to increase their risk profile and are more likely to be rescued after the crisis. Likewise, we find
that as competition becomes stronger, increased capital levels do not buffer the direct impact of
competition on bank risk. While the latter finding most likely reflects the regulators’ tendency to demand
higher capital in a more competitive environment, it also shows that higher levels of bank do not prove to
be sufficient to ensure financial stability as competition increases. As in Gale (2010) and Beck (2008) our
results suggests that, first, excessive reliance on bank capital does not seem enough to ensure financial
stability. Second they suggest that supervisory authorities should cooperate closely with competition
15
Results available upon request. 16
We check the consistency of our results using an alternative continuous variable accounting for bank risk as perceived by the
markets: the expected default frequency of each bank. Results available upon request.
authorities in particular when evaluating the financial stability implications of bank capital regulation or
securitization.
References
Ahn, J.-H. and R. Breton (2011), “Securitization, competition and incentive to Monitor,” mimeo.
Ai, C. and E. Norton (2003), “Interaction terms in Logit and Probit models,” Economics Letters 80, pp. 123-129.
Albertazzi, U., G. Eramo, L. Gambacorta and C. Salleo (2011), “Securitization is not that evil after all,” Bank for
International Settlements Working Papers 341.
Allen, F. (1990), “The market for information and the origin of financial intermediation,” Journal of Financial
Intermediation 1, pp. 3-30.
Allen, F. and D. Gale (2000), Comparing financial systems, Cambridge, MA: MIT Press.
Allen, F. and D. Gale (2004), “Competition and financial stability,” Journal of Money, Credit and Banking 36(3),
pp. 453-80.
Altunbas, Y., S. Manganelli and D. Marques-Ibanez (2011), “Bank risk during the financial crisis: do business
models matter”, European Central Bank working paper 1394.
Beck, T. (2008), “Bank competition and financial stability: friends or foes?,” The World Bank Policy Research
Working Paper Series 4656.
Beltratti, A. and R.M. Stulz (2012), “Why did some banks perform better during the credit crisis? A cross-country
study of the impact of governance and regulation,” Journal of Financial Economics 105 (1), pp. 1-17.
Benmelech, E., J. Dlugosz and V. Ivashina (2012), “Securitization without adverse selection: The case of CLOs,”
Journal of Financial Economics 106(1), pp. 91-113.
Benston, G. (1998), Regulating Financial Markets: A Critique and Some Proposals,
Institute of Economic Affairs, Hobart paper, 135.
Berger, A. and C. Bouwman (2013), “How does capital affect bank performance during financial crises?,” Journal
of Financial Economics, forthcoming.
Boone, J. (2008), “A new way to measure competition,” Economic Journal 118, 1245 -1261.
Boyd, J.H., and E.C. Prescott (1986), “Financial intermediary-coalitions,” Journal of Economic Theory 38, pp. 211-
32.
Boyd, J.H., and G. De Nicoló (2005), “The theory of bank risk-taking and competition revisited,” Journal of
Finance 60, pp.1329-343.
Carletti, E. and P. Hartmann (2003), “Competition and stability: what's special about banking?,” European Central
Bank Working Paper Series 146.
Cebenoyan, A. and P. Strahan (2004), “Risk management, capital structure and lending at banks,” Journal of
Banking and Finance 28, pp. 19-43.
Cetorelli, N. (2001), “Competition among banks: good or bad?,” Economic Perspectives, Federal Reserve Bank of
Chicago, issue Q I, pp. 38-48.
Corts, K. (1999), “Conduct parameters and the measurement of market power,” Journal of Econometrics
88, pp. 227-250.
Delis, M.D. and P.K. Staikouras (2011), “Supervisory effectiveness and bank risk,” Review of Finance, 15(3), pp.
511-543.
Demirguc-Kunt, A., E. Detragiache, and O. Merrouche (2010), “Bank capital: lessons from the financial
crisis,” The World Bank Policy Research Working Paper Series 5473.
Gale, D. and O. Özgür (2005), “Are bank capital ratios too high or too low? Risk aversion, incomplete markets, and
optimal capital structure,” Journal of the European Economic Association 3(2-3), pp. 690-700.
Gale, D. (2010), “The effects of bank capital on lending: What do we know, and what does it mean?,” International
Journal of Central Banking 6(34), pp. 187-204,
Genesove, D. and W.P. Mullin, (1998), “Testing static oligopoly models: conduct and cost in the sugar industry,
1890-1914,” Rand Journal of Economics 29, 355- 377.
Gorton, G. and G. Ordonez (2012), “Collateral crises,” NBER Working Papers 17771.
Gorton, G.B. and A. Metrick (2012), “Securitization”, The Handbook of the Economics of Finance, ed. G.
Constantinides, M. Harris, and R. Stulz.
Graddy, K. (1995), “Testing for imperfect competition of the Fulton fish market,” RAND Journal of Economics 26,
pp. 75-92.
Greene, W. (2010), “Testing hypotheses about interaction terms in nonlinear models,” Economics Letters 107(2),
pp. 291-296.
Haldane, A and V. Madouros (2012), “The Dog and the Frisbee”, speech presented at the Federal Reserve Bank of
Kansas City’s 36th economic policy symposium “The Changing Policy Landscape”, Jackson Hole,
Wyoming.
Hellmann, F., K.C. Murdock and J. Stiglitz, (2000), “Liberalization, moral hazard in banking and prudential
regulation: are capital requirements enough?,” American Economic Review 90(1), pp. 147 -165.
Keeley, M.C. (1990), “Deposit insurance, risk and market power in banking,” American Economic Review 80, pp.
1183-200.
Keys, B., T. Mukherjee, A. Seru and V. Vig (2010), “Did Securitization Lead to Lax Screening? Evidence from
Subprime Loans”, Quarterly Journal of Economics 125, pp. 307-362.
Marques-Ibanez, D. and M. Scheicher (2010), “Securitisation: Causes and consequences,” in A. Berger, P.
Molyneux and J. Wilson (eds), Handbook of Banking, Oxford University Press pp. 599-633.
Martinez-Miera, D. and R. Repullo (2010), “Does competition reduce the risk of bank failure?,” Review of
Financial Studies 23(10), pp. 3638-3664.
Matutes, C. and X. Vives (1996), “Competition for deposits, fragility, and insurance,” Journal of Financial
Intermediation 5(2), pp. 184-216.
Mian, A. and A. Sufi (2009), “The consequences of mortgage credit expansion: evidence from the U.S. mortgage
default crisis,” Quarterly Journal of Economics 124, pp. 1449-1496.
Nijskens, R. and W. Wagner, (2011), “Credit risk transfer activities and systemic risk: How banks became less
risky individually but posed greater risks to the financial system at the same time,” Journal of Banking and
Finance 35(6), pp. 1391-1398.
Ramakrishnan, R., and A.V. Thakor (1984), “Information reliability and a theory of financial intermediation,”
Review of Economic Studies 51, pp. 415-32.
Stein, J.C. (2011), “Monetary policy as financial stability regulation,” NBER Working Papers 16883.
VanHoose, D. (2007), “ Theories of bank behaviour under capital regulation,” Journal of Banking and
Finance 31, pp. 3680-3697.
Van Leuvensteijn, M., J.A. Bikker, A.A.R.J.M. van Rixtel and C. Kok Sorensen (2011), “A new approach to
measuring competition in the loan markets of the euro area,” Applied Economics, 3155-3167.
Van Leuvensteijn, M., C. Kok Sorensen, J.A. Bikker and A.A.R.J.M. van Rixtel (2013), “Impact of bank
competition on the interest rate pass-through in the euro area,” Applied Economics, 1359-1380.
Vives, X. (2001), “Competition in the changing world of banking,” Oxford Review of Economic Policy, 17(4), pp.
535-547.
Vives, X. (2010), “Competition and stability in banking,” CESifo Working Paper Series 3050.
Williamson, S.D. (1986), “Costly monitoring, loan contracts and equilibrium credit rationing,” Quarterly Journal
of Economics 102, pp. 135-46.
Table I
DESCRIPTIVE STATISTICS This table presents the names of the variables employed in our empirical analysis, indicating the data sources, as well as a brief description of how the variables have been
constructed. More detailed information, plus all publicly available data, are available upon request. This table presents the summary statistics of the variables used in our
paper (see Section II for further details). Unless stated otherwise, descriptive statistics are derived from the average values calculated on the basis of quarterly data for the
pre-crisis or the crisis period. The variable accounting for bank risk is calculated during the crisis period (2007Q4 to 2009Q4). The variables accounting for securitization,
capital structure, size, loan growth and deposits are calculated from the averages of quarterly data for individual banks for the pre-crisis period (2003Q4 to 2007Q3). GDP
growth and competition variables are calculated from the country using quarterly and yearly data respectively averaged over the pre-crisis period already mentioned.
183.3430.67822.73747.37245.985495COMP*CAP
54.9010.0003.3520.0000.508495COMP*SEC
1.8460.5850.2031.3411.277495Quarterly changes in real GDP during
the pre-crisis period (2003Q4 to
2007Q3)
Bank for International SettlementsGDP GROWTH
89.8257.05216.36273.34669.049495Short-term demand deposits to total
assets * 100 during the pre-crisis
period (2003Q4 to 2007Q3)
BloombergDEP
13.270-0.2742.4026.0726.698495Individual bank lending growth
minus the average loan growth of all
banks over a specific quarter during
the pre-crisis period (2003Q4 to
2007Q3)
Authors' calculationEXLEND
13.9803.9702.1406.9627.657495Logarithm of total assets (USD
millions) during the pre-crisis period
(2003Q4 to 2007Q3)
BloombergSIZE
32.7401.4243.4778.7219.011495Tier I capital to total assets * 100
during the pre-crisis period (2003Q4
to 2007Q3)
BloombergCAP
9.8040.0000.7060.0000.133495Ratio of total securitization to total
assets * 100 during the pre-crisis
period (2003Q4 to 2007Q3)
DCM Analytics DealogicSEC
5.6000.2341.4255.6004.970495The average relationship between
market share and marginal costs
estimated by country-level by year
for the period 1994-2004, the Boone
indicator
Authors' calculationCOMP
1.0000.0000.4600.0000.303495Binary variable – with a value of 1 if
public financial support was received
during the crisis period (2007Q4 to
2009Q4) and 0, if otherwise
European Commission, central
banks, Bank for International
Settlements, governmental
institutions and Bloomberg
RISKY
Max.Min.St. Dev.MedianAverageNDescriptionSourceVariable
183.3430.67822.73747.37245.985495COMP*CAP
54.9010.0003.3520.0000.508495COMP*SEC
1.8460.5850.2031.3411.277495Quarterly changes in real GDP during
the pre-crisis period (2003Q4 to
2007Q3)
Bank for International SettlementsGDP GROWTH
89.8257.05216.36273.34669.049495Short-term demand deposits to total
assets * 100 during the pre-crisis
period (2003Q4 to 2007Q3)
BloombergDEP
13.270-0.2742.4026.0726.698495Individual bank lending growth
minus the average loan growth of all
banks over a specific quarter during
the pre-crisis period (2003Q4 to
2007Q3)
Authors' calculationEXLEND
13.9803.9702.1406.9627.657495Logarithm of total assets (USD
millions) during the pre-crisis period
(2003Q4 to 2007Q3)
BloombergSIZE
32.7401.4243.4778.7219.011495Tier I capital to total assets * 100
during the pre-crisis period (2003Q4
to 2007Q3)
BloombergCAP
9.8040.0000.7060.0000.133495Ratio of total securitization to total
assets * 100 during the pre-crisis
period (2003Q4 to 2007Q3)
DCM Analytics DealogicSEC
5.6000.2341.4255.6004.970495The average relationship between
market share and marginal costs
estimated by country-level by year
for the period 1994-2004, the Boone
indicator
Authors' calculationCOMP
1.0000.0000.4600.0000.303495Binary variable – with a value of 1 if
public financial support was received
during the crisis period (2007Q4 to
2009Q4) and 0, if otherwise
European Commission, central
banks, Bank for International
Settlements, governmental
institutions and Bloomberg
RISKY
Max.Min.St. Dev.MedianAverageNDescriptionSourceVariable
Table II
PROBIT ESTIMATES OF THE PROBABILITY OF BEING RESCUED This table provides the probit estimates of seven specifications on the likelihood of receiving financial support. The first column shows the effect of bank balance sheet
variables and competition. Column (II) shows these effects with country dummies. Column (III) introduces GDP growth as macro variable. Columns (V) and (VII)
introduce the interaction effects, while Columns (IV) and (VI) show the interaction effects without the macro variable GDP growth (see Section II for further details and
Table I for variable definitions). The dependent variable is calculated during the crisis period (2007Q3 to 2009Q4). Regressors are calculated as averages of quarterly
data for individual banks during the pre-crisis period (2003Q4 to 2007Q4) unless otherwise indicated. *, ** and *** indicate statistical significance at the 10%, 5% and
1% levels, respectively.
0.59450.42230.52920.56330.15880.16390.2822Hosmer–Lemeshow test p-value
6.478.127.076.7511.8311.729.76Hosmer–Lemeshow test
73.9773.673.173.3573.2374.673.23Percent correctly classified
54.88/76.1453.01/75.9750.63/75.5551.85/75.7651.28/75.5957.14/76.8051.25/75.66Percent true positives/negatives
0.14270.14850.14330.1410.13790.16860.1372R2
495495495495495495495No. of observations
(1.3304)(1.3052)(0.7225)(0.6618)(0.6947)(0.8844)(0.6397)
-1.0530-1.0233***-3.5494***-3.1943***-3.9954***-5.6402***-3.7414CONSTANT
(0.0116)(0.0112)
***0.0368***0.0308COMP*CAP
(0.0811)(0.0787)
**0.1825**0.1706COMP*SEC
Competition interactions
YesCOUNTRY DUMMIES
(0.1303)(0.1473)(0.1421)
*0.22900.15810.1064GDP GROWTH
Macro Economic variables
(0.0030)(0.0032)(0.0033)(0.0033)(0.0035)(0.0034)(0.0035)
***-0.0087***-0.0089***-0.0091***-0.0090**-0.0088**-0.0080**-0.0088DEP
(0.0242)(0.0257)(0.0261)(0.0264)(0.0268)(0.0284)(0.0270)
0.02990.03190.03270.03410.02990.03460.0310EXLEND
(0.0308)(0.0327)(0.0319)(0.0325)(0.0330)(0.0284)(0.0333)
0.04470.0457*0.05290.0513*0.0620**0.0580*0.0605SIZE
(0.0629)(0.0613)(0.0072)(0.0070)(0.0073)(0.0078)(0.0071)
***-0.2217***-0.1885***-0.0248***-0.0238***-0.0230***-0.0283***-0.0224CAP
(0.0507)(0.0388)(0.3770)(0.3623)(0.0668)(0.0351)(0.0557)
**-0.1008**-0.0923**-0.9549**-0.8783**-0.1470***-0.1044**-0.1371SEC
(0.0593)(0.0574)(0.0298)(0.0254)(0.0300)(0.0388)(0.0251)
-0.0794-0.0279***0.0841***0.1017***0.11530.0615***0.1260COMP
(VII) (VI) (V) (IV) (III) (II) (I)
0.59450.42230.52920.56330.15880.16390.2822Hosmer–Lemeshow test p-value
6.478.127.076.7511.8311.729.76Hosmer–Lemeshow test
73.9773.673.173.3573.2374.673.23Percent correctly classified
54.88/76.1453.01/75.9750.63/75.5551.85/75.7651.28/75.5957.14/76.8051.25/75.66Percent true positives/negatives
0.14270.14850.14330.1410.13790.16860.1372R2
495495495495495495495No. of observations
(1.3304)(1.3052)(0.7225)(0.6618)(0.6947)(0.8844)(0.6397)
-1.0530-1.0233***-3.5494***-3.1943***-3.9954***-5.6402***-3.7414CONSTANT
(0.0116)(0.0112)
***0.0368***0.0308COMP*CAP
(0.0811)(0.0787)
**0.1825**0.1706COMP*SEC
Competition interactions
YesCOUNTRY DUMMIES
(0.1303)(0.1473)(0.1421)
*0.22900.15810.1064GDP GROWTH
Macro Economic variables
(0.0030)(0.0032)(0.0033)(0.0033)(0.0035)(0.0034)(0.0035)
***-0.0087***-0.0089***-0.0091***-0.0090**-0.0088**-0.0080**-0.0088DEP
(0.0242)(0.0257)(0.0261)(0.0264)(0.0268)(0.0284)(0.0270)
0.02990.03190.03270.03410.02990.03460.0310EXLEND
(0.0308)(0.0327)(0.0319)(0.0325)(0.0330)(0.0284)(0.0333)
0.04470.0457*0.05290.0513*0.0620**0.0580*0.0605SIZE
(0.0629)(0.0613)(0.0072)(0.0070)(0.0073)(0.0078)(0.0071)
***-0.2217***-0.1885***-0.0248***-0.0238***-0.0230***-0.0283***-0.0224CAP
(0.0507)(0.0388)(0.3770)(0.3623)(0.0668)(0.0351)(0.0557)
**-0.1008**-0.0923**-0.9549**-0.8783**-0.1470***-0.1044**-0.1371SEC
(0.0593)(0.0574)(0.0298)(0.0254)(0.0300)(0.0388)(0.0251)
-0.0794-0.0279***0.0841***0.1017***0.11530.0615***0.1260COMP
(VII) (VI) (V) (IV) (III) (II) (I)
Chart I
MARGINAL EFFECTS In Panel A (Panel B), the red line plots the expected marginal effect based on the estimates of the interaction effects between competition and securitization
(capital) as reported in Table II (columns V and VII) over the sample population. The dots report the corrected values of the marginal effect of this interaction term
varying over the sample population on the predicted risk of financial support.
Chart II
Z-STATISTIC AS A FUNCTION OF THE PREDICTED PROBABILITY In Panel A (Panel B), the blue dots reflect the z-values of the correct interaction effect reported in Panel A (Panel B) in Chart I for the interaction effects between
competition and securitization (capital) as reported in Table II (columns V and VII) over the sample population.
Panel A: Interaction effect of competition and securitization
Panel B: Interaction effect of competition and capitalisation
Panel A: Interaction effect of competition and securitization
Panel B: Interaction effect of competition and capitalisation